CN110825833B - Method for predicting user moving track point - Google Patents
Method for predicting user moving track point Download PDFInfo
- Publication number
- CN110825833B CN110825833B CN201911096060.7A CN201911096060A CN110825833B CN 110825833 B CN110825833 B CN 110825833B CN 201911096060 A CN201911096060 A CN 201911096060A CN 110825833 B CN110825833 B CN 110825833B
- Authority
- CN
- China
- Prior art keywords
- point
- points
- time
- historical track
- dwell
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/29—Geographical information databases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0202—Market predictions or forecasting for commercial activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/14—Travel agencies
Abstract
A method for predicting track points of a user at a target moment comprises the following steps: performing density clustering on a plurality of historical track points in a period of time in the past of a user, and obtaining at least one resident point based on the density clustering result; selecting a resident point meeting a preset condition with the target moment from the at least one resident point as a candidate resident point; and based on a predetermined rule, selecting one dwell point from the candidate dwell points as the predicted track point.
Description
Technical Field
The present invention relates to mobile communication, and more particularly, to a method for predicting a user's moving track point in a mobile communication system.
Background
In recent years, in a mobile communication system, predicting the movement of a user or a user of a portable mobile terminal based on a history track point has become a new hotspot in the field of movement track prediction. The concept of historical track points is well known in the field of mobile communications. The historical track points are mainly obtained by means of data observed by a sensor embedded in the portable mobile terminal and data recorded by a base station signal connected with a user, and are used for representing the position of the user at a certain moment in the past. Analyzing and mining the historical track points and constructing a prediction model can be helpful for scientific decision making in the application of the related field. The prediction of the movement track generally refers to predicting a place to be reached by a user after a short period of time in the future according to historical track points and information such as the current time and the position of the user. For example, a travel agency may schedule marketing strategies for travel products through the results of the prediction of the movement trajectories of potential customers.
The prior art proposes a method for predicting a moving object trajectory end point, which employs a bayesian model and a markov model. However, this method does not eliminate the influence of the non-aftereffect of the markov chain on the prediction accuracy, i.e. the user historical track points are not fully utilized, resulting in low accuracy. The prior art also provides a location sequence prediction method based on the historical track points of the user. The method calibrates the staying places, and calculates the distribution of a plurality of places at the next moment by dividing the track of the user into a place visiting time sequence and combining the current situation characteristics of the user. And then, weighting by artificially defining and adjusting the weight, and finally predicting the location distribution of the user at the next moment. However, the method does not consider the regularity implicit in the daily behaviors of human beings, and also adds artificial adjustment factors, which can influence the objectivity and accuracy of the track prediction result.
Disclosure of Invention
Aiming at the problems in the prior art, the application aims to provide a user moving track point prediction method which can make full use of the user historical track points and has higher accuracy.
A method for predicting a predicted track point of a user at a target moment comprises the following steps: performing density clustering on a plurality of historical track points in a period of time past by a user, and obtaining at least one resident point based on a density clustering result; selecting a resident point meeting a preset condition with a target moment from the at least one resident point as a candidate resident point; and selecting one resident point from the candidate resident points as the predicted track point based on a predetermined rule.
The method for predicting the user moving track points is based on the prediction of the density clustering result of the user historical track points, and makes full use of the user historical track points, so that the prediction result is more accurate.
Drawings
Fig. 1 shows a flowchart of a method of predicting a movement trajectory of a user according to an embodiment of the present invention.
Fig. 2 shows a flowchart of a method of predicting a movement trajectory of a user according to another embodiment of the present invention.
Detailed Description
The content of the invention will now be discussed with reference to a number of exemplary embodiments. It is to be understood that these examples are discussed only to enable those of ordinary skill in the art to better understand and thus implement the teachings of the present invention, and are not meant to imply any limitations on the scope of the invention.
As used herein, the term "include" and its variants are to be read as open-ended terms meaning "including, but not limited to. The term "based on" is to be read as "based, at least in part, on". The terms "one embodiment" and "an embodiment" are to be read as "at least one embodiment". The term "another embodiment" is to be read as "at least one other embodiment".
Referring to fig. 1, a flow chart of a method of predicting a track point (predicted track point) of a user at a target time in the future according to one embodiment of the present invention is shown. FIG. 2 illustrates in more detail a flow chart of a method of predicting a user movement trajectory of another embodiment.
The steps will be described in detail with reference to fig. 1 and 2.
Step 1: performing density clustering on a plurality of historical track points in a period of time in the past of a user, based on the density clustering Resulting in at least one dwell point。
According to the embodiment of the invention, historical data of communication between the user and the base station can be obtained from the base station, and the position information of the user at a plurality of past moments can be obtained from the data, so that the historical track points of the user can be obtained. For example, the history track point contains time information and position information represented by a 7-bit GeoHash position (simply referred to as "geo 7"). The GeoHash is a spatial index technology for converting two-dimensional longitude and latitude position data into a one-dimensional character string. The longer the character string, the smaller and more accurate the range of representation; the smaller the length of the character string, the wider the range of representation, and the more similar the character string, the closer the representation distance.
In one embodiment, the Density Clustering employed by the method may be the commonly used DBSCAN (Density-Based Spatial Clustering of Applications with Noise) Density Clustering. DBSCAN is a density-based clustering algorithm that generally assumes that classes can be determined by how closely the samples are distributed. Samples of the same class are closely related, i.e., samples of the same class must exist a short distance around any sample of the class. By classifying closely connected samples into one class, a cluster class is obtained. By classifying all groups of closely connected samples into different categories, we obtain the final results of all the clustering categories. By performing DBSCAN density clustering on geo7 of the historical track points, at least one (typically multiple) cluster center and cluster can be obtained. In one embodiment, the cluster center may be the user's residence point. The dwell point may be used to represent an approximate location point where the user stays for a long time. Since the dwell point is computed by the clustering algorithm, it is likely not any "trace point". The "cluster centers" and "dwell points" contain no time information, only location information (i.e., the GeoHash value). A cluster is a subset of all the historical track points for which the density cluster is directed. Each cluster corresponds to one cluster center, and the historical track points in each cluster can be understood as the historical track points generated by the activity of the user near the residence points. Although DBSCAN density clustering is employed in the present embodiment, it should be understood that in other embodiments, other density clustering algorithms may be employed.
Replenishment of stagnation points
Based on special or extreme considerations, in one embodiment of the invention, the dwell points may be supplemented based on predetermined dwell point supplement rules. And then combining the residence points obtained by density clustering in the step 1 to obtain residence points with more sufficient quantity. The following three dwell point replenishment rules are provided in this application:
(1) supplemental rules to relax density clustering conditions
Since base stations near a country or a remote location are sparse and far away, when a density clustering algorithm is used to obtain clustering centers, it may be difficult to obtain the required number of clustering centers due to too strict density clustering conditions. The density clustering condition is expressed as a setting of a parameter, which may include, for example, a cluster radius, a number within a cluster, and the like. Thus, in one embodiment, the setting of the correlation parameter may be relaxed appropriately, for example, to make the cluster radius larger or to reduce the minimum number within the cluster to obtain the desired number of cluster centers.
However, simply relaxing all conditions may result in a large amount of normal data identifying residents that are too dense to be connected together, thereby losing the referential meaning of the supplemental residents. Therefore, this rule is typically only employed when the original density cluster does not result in a stagnation point.
(2) Complementary rules based on time and distance thresholds
In one embodiment, the dwell point may also be supplemented with reference to both time and distance. Specifically, in the historical track points of the user, if the time difference between any two historical track points is greater than a certain threshold value and the distance difference is smaller than a certain threshold value, the geo7 of any one of the two historical track points is marked as a staying point. For example, a work of a person requires shutdown at work time, so that the time difference between track points before and after shutdown is large and the distance between the track points after the shutdown is small, the time threshold value can be set to be 4 hours, the distance threshold value is set to be 500 meters, and the historical track points meeting the conditions are set to be the stay points. This approach is suitable for situations where the user is using the handset less often resulting in sparse anchor points.
(3) Frequency statistics based supplemental rules
In the case that the historical track points are extremely sparse, or when all other logics do not acquire the information of any resident point, statistical information based on the frequency of occurrence of the historical track points can be adopted to supplement the resident points. This is often the case in very remote countries where the user's handset can only connect to a few base stations throughout the year and where the base stations are located at great distances. For this case, historical track points that occur more frequently (above a certain threshold) can be marked as dwell points.
The above dwell point supplement rules may be used alone or in combination. For example, in one embodiment, the above supplemental rules may be used in combination as follows:
firstly, dividing the historical track points of the user into track points of a plurality of days according to the days, respectively carrying out density clustering on the track points of each day by using the density clustering model in the step 1, and firstly using conventional parameter clustering.
And in the case that the resident point is not obtained, supplementing the resident point by using a supplement rule for relaxing the density clustering condition, and simultaneously supplementing the resident point by using a supplement rule based on time and distance threshold values.
If the residence point is not obtained, supplementing the residence point by adopting a supplement rule based on frequency statistics; and finally, performing density clustering merging on the obtained multi-day residence points (namely, overlapping and de-duplicating all residence points of the multi-day residence points) to obtain a sufficient number of residence points.
Step 2: selecting the parking points meeting the preset condition with the target time from the at least one parking point obtained in the step 1
The stay point is used as a candidate stay point.
The target time instant represents the point in time to be predicted. The target time may be a time some time after the current time, such as a time after 0.5 hours, 1 hour, or 2 hours, etc. For example, if the current time is 15:00 PM and the location of the target user is predicted after 2 hours, the target time is 17:00 PM. The candidate stay points represent points at which the user may stay at the target time, which may include multiple stay points. Candidate dwell points are only possible points of the preliminary screening. The final predicted point will be generated from the candidate resident points. In the embodiment of the present invention, a dwell point satisfying a predetermined condition with respect to the target time may be selected as a candidate dwell point from the dwell points obtained in step 1.
In one embodiment, the candidate anchor point may be selected by:
establishing a mapping relation between each resident point in the at least one resident point and the corresponding historical track point;
selecting historical track points within a time range near the target time; and
and acquiring the resident points corresponding to the selected historical track points as candidate resident points based on the mapping relation.
The mapping relation between the user historical track point and the residence point can be established according to the acquisition mode of the residence point.
For the resident points obtained by density clustering, each resident point corresponds to one clustering cluster, all historical track points in the clustering clusters correspond to the resident points, and a mapping relation can be established according to the resident points. For a dwell point obtained based on a time and distance threshold supplementary rule, a mapping relation can be established between two historical track points obtaining the dwell point and the dwell point. For the resident points obtained based on the frequency statistics and the supplement rules, the mapping relation can be directly established between the historical track points corresponding to the obtained resident points and the resident points.
In one embodiment, historical track points may be marked with information of the resident points during the process of establishing the mapping relationship. For example, all the obtained residence points may be numbered 1, 2, 3, called residence point IDs, and then the historical track points that conform to the mapping relationship are marked with the residence point IDs, and finally, the historical track points partially marked with the residence point IDs are obtained.
Specifically, for the resident points obtained by density clustering, each resident point corresponds to a cluster, and all historical track points in the cluster mark the ID of the resident point. For the resident points obtained based on the time and distance threshold value supplement rules, obtaining the IDs of the two historical track points of the resident points for marking the resident points; and for the resident points obtained based on the frequency statistics and supplement rules, marking the ID of the resident points on the historical track points corresponding to the obtained resident points.
Next, historical track points may be selected within a predetermined time range around the target time. For example, the plurality of history track points may be selected in a time range of 1 hour (i.e., the predetermined time range is 2 hours), 0.5 hour (i.e., the predetermined time range is 1 hour), 10 minutes (i.e., the predetermined time range is 20 minutes), or the like around the target time. The predetermined time range is typically less than the time difference between the target time and the current time to make the prediction more accurate. Preferably, the predetermined time range is less than half of the time difference between the target time and the current time. For example, if the user's position 2 hours later is to be predicted (i.e., the time difference between the target time and the current time is 2 hours), the time range is preferably within 1 hour before and after the target time, for example, 0.5 hour before and after, 10 minutes before and after, and the like. And then according to the established mapping relation, finding out corresponding resident points from the track points to serve as candidate resident points. For example, the residing point can be found according to the mark, and the points marked by the residing point ID can be directly selected.
According to one embodiment of the invention, a more preferable method for selecting historical track points in a time range around a target time is provided. The method comprises the following steps:
obtaining the current time T0The corresponding time T of each historical track point in a first plurality of historical track points adjacent to the current track point within a previous preset time rangei；
Obtaining the closest in time to Ti+T1-T0The second plurality of historical track points as the selected historical track points, wherein T is1Representing the target time instant.
The current time T0The preceding and following predetermined time ranges may be different from the predetermined time range around the above-described target time. Similarly, this current time T0The previous and subsequent predetermined time ranges are typically less than the time difference between the target time and the current time to make the prediction more accurate. Preferably, the predetermined time range is less than half of the time difference between the target time and the current time.
The following is a detailed description by way of example.
The known data is the current T of the user0Pushing the historical track points of the user within 30 days forward, and predicting the user two hours later (T)0+2h time, i.e. target time T1) Position of (2), note T0The track point of the moment is p0Then, the embodiment of the present invention may find a candidate staying point according to the staying point candidate model based on the history track matching, and determine the final predicted position by the distance.
The method can be specifically carried out as follows:
(1) find the first 30 days T0H to T0The + h period (i.e. theT0One hour before and after time) and with T0Recording all history track points adjacent to geo7 of the moment track point as P, and acquiring time T of all history track points in the set Pi。
Here, "adjacent" means: at T0The geo7 of the time track point extends into the squared figure.
geo7 is an area of approximately 152 meters x 152 meters, and the Sudoku includes T0The point of the moment track is geo7 and the 8 equally sized regions adjacent to it (geo 7). The expression of the trace point is geo7, and it is understood that this region represents the trace point.
See the following figure, wherein the middle gray background region represents T0The point of time trace geo 7. And T0Historical track point and T in 8 white background areas (geo7) of the same size around geo7 of the time of day track point0The relationship of the time of day trace points is "adjacent" as described herein.
| geo7 | geo7 | geo7 |
| geo7 | geo7 | geo7 |
| geo7 | geo7 | geo7 |
It should be understood that although the time period selected here is T0One hour before and after the moment, howeverOther time can be selected according to actual conditions and needs, for example, T can be used00.5h, 10 minutes and the like before and after the moment.
For example, if the current time is 15:00 pm and the location of the target user is predicted after 2 hours (i.e., 17:00), all historical track points adjacent to the current time track point can be found in a time period of 14:00 to 16: 00. Time T of these trace pointsiMay vary, for example, the time may be 14:28, 15:01, 15:59, and so forth.
(2) The time T of all historical track points in P is calculatediPush back for 2 hours (i.e., T)1-T0) Find the closest T in timei+T1-T0(i.e., T)i+2h) as P _2h, and the time of the historical track point in P _2h as Th。
For example, for a point with time 14:28, push back for 2 hours, i.e., 16:28, find the historical track point closest to that time 16: 28. And (3) finding out the closest point in time 2 hours later for all the adjacent points found in the step (1), wherein the set is P _2 h.
Next, the historical track points marked with the stay point ID in P _2h, marked as Pr _2h, and all the stay points marked in Pr _2h, marked as R, can be found, called candidate stay points.
And step 3: selecting one dwell point from the candidate dwell points as a prediction of a target time instant based on a predetermined rule
And (5) tracing points.
The predetermined rule may be, for example, selecting the dwell point that is closest to the user's current track point. For example, calculate T0Time trace point p0Distance H from all the points of residence in RRTaking the residence point with the shortest distance as the user T0Predicted position at time +2 h. In one embodiment, if none of the historical track points in P _2h mark the resident point ID, then the distance P in P _2h is added0Recent historical track points as user T0Predicted position at time +2 h.
In one embodiment, the reservation is madeThe rule may be based on a combination of time of day and distance. For example, H may be usedRTime T multiplied by historical track point in P _2hhAnd T0Absolute value of difference of +2h (i.e., target time) (i.e., | T)0+2h-Th|), the candidate dwell point with the smallest product is selected.
The method and apparatus of the embodiments of the present invention may be implemented as a pure software module (for example, a software program written in Python language), as a pure hardware module (for example, a special ASIC chip or an FPGA chip) as required, or as a module combining software and hardware (for example, a firmware system storing fixed codes).
Another aspect of the invention is a computer-readable medium having computer-readable instructions stored thereon that, when executed, perform a method of embodiments of the invention.
It will be appreciated by persons skilled in the art that the foregoing description is only exemplary of the invention and is not intended to limit the invention. The present invention may include various modifications and variations. Any modifications and variations within the spirit and scope of the present invention should be included within the scope of the present invention.
Claims (8)
1. A method for predicting a predicted track point of a user at a target moment comprises the following steps:
performing density clustering on a plurality of historical track points in a period of time in the past of a user, and obtaining at least one resident point based on the density clustering result;
selecting a resident point meeting a preset condition with the target moment from the at least one resident point as a candidate resident point; and
selecting one dwell point from the candidate dwell points as the predicted trajectory point based on a predetermined rule,
wherein, selecting the dwell point which meets the preset condition with the target time from the at least one dwell point as the candidate dwell point comprises:
establishing a mapping relation between each resident point in the at least one resident point and the corresponding historical track point,
selecting historical track points within a first predetermined time range around the target time, an
Based on the mapping relation, obtaining the dwell point corresponding to the selected historical track point as a candidate dwell point,
and wherein selecting historical track points within a first predetermined time range around the target time comprises:
obtaining the current time T0The corresponding time T of each historical track point in a first plurality of historical track points adjacent to the current track point in a second preset time rangeiAnd an
Obtaining the closest in time to Ti+T1-T0The second plurality of historical track points as the selected historical track points, wherein T is1Representing the target time instant.
2. The method of claim 1, wherein the obtaining at least one dwell point based on the results of density clustering comprises treating a cluster center of the density clustering as a dwell point.
3. The method of claim 2, wherein the obtaining at least one dwell point based on the results of density clustering further comprises: the additional dwell points are supplemented based on predetermined rules.
4. The method of claim 3, wherein supplementing additional dwell points based on predetermined rules comprises a combination of one or more of:
relaxing the condition of density clustering to obtain more clustering centers as the residence points;
based on time and distance thresholds, any one of two historical track points of which the time difference is greater than a certain threshold and the distance difference is less than a certain threshold is marked as a resident point;
based on frequency statistics, historical track points with frequency higher than a certain threshold are used as residence points.
5. The method of claim 1, wherein the first predetermined time range is less than half of the time difference between the target time and the current time.
6. A method according to claim 1, wherein the predetermined rule is to select one of the candidate resident points that is closest to the user's current track point as the predicted track point.
7. The method of claim 1, wherein the target time is a time 0.5 hours, 1 hour, or 2 hours after the current time.
8. A computer-readable medium having stored thereon computer-executable instructions capable, when executed by a computer, of performing the method of any one of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911096060.7A CN110825833B (en) | 2019-11-11 | 2019-11-11 | Method for predicting user moving track point |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911096060.7A CN110825833B (en) | 2019-11-11 | 2019-11-11 | Method for predicting user moving track point |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110825833A CN110825833A (en) | 2020-02-21 |
| CN110825833B true CN110825833B (en) | 2022-05-17 |
Family
ID=69553937
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911096060.7A Active CN110825833B (en) | 2019-11-11 | 2019-11-11 | Method for predicting user moving track point |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110825833B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111506691A (en) * | 2020-04-20 | 2020-08-07 | 杭州数澜科技有限公司 | Track matching method and system based on depth matching model |
| CN111553732B (en) * | 2020-04-24 | 2024-02-23 | 北京爱笔科技有限公司 | Method and device for processing movement track |
| CN111698332A (en) * | 2020-06-23 | 2020-09-22 | 深圳壹账通智能科技有限公司 | Method, device and equipment for distributing business objects and storage medium |
| CN111831178B (en) * | 2020-06-29 | 2023-01-17 | 中国科学院软件研究所 | Method and system for assisting target selection in three-dimensional environment based on motion trend information |
| CN113486927B (en) * | 2021-06-15 | 2024-03-01 | 北京大学 | Priori probability-based unsupervised track access place labeling method |
| CN114265912A (en) * | 2021-12-27 | 2022-04-01 | 智慧足迹数据科技有限公司 | Residence point generation method and related device |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105409306A (en) * | 2014-06-11 | 2016-03-16 | 华为技术有限公司 | Method and apparatus for predicting location of mobile terminal |
| CN109413587A (en) * | 2018-09-20 | 2019-03-01 | 广州纳斯威尔信息技术有限公司 | User trajectory prediction technique based on WiFi log |
| CN110275531A (en) * | 2019-06-21 | 2019-09-24 | 北京三快在线科技有限公司 | The trajectory predictions method, apparatus and unmanned equipment of barrier |
| CN110322067A (en) * | 2019-07-03 | 2019-10-11 | 中山大学 | Location of mobile users prediction technique based on factor graph model |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8577815B2 (en) * | 2009-10-23 | 2013-11-05 | GM Global Technology Operations LLC | Method and system for concurrent event forecasting |
-
2019
- 2019-11-11 CN CN201911096060.7A patent/CN110825833B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105409306A (en) * | 2014-06-11 | 2016-03-16 | 华为技术有限公司 | Method and apparatus for predicting location of mobile terminal |
| CN109413587A (en) * | 2018-09-20 | 2019-03-01 | 广州纳斯威尔信息技术有限公司 | User trajectory prediction technique based on WiFi log |
| CN110275531A (en) * | 2019-06-21 | 2019-09-24 | 北京三快在线科技有限公司 | The trajectory predictions method, apparatus and unmanned equipment of barrier |
| CN110322067A (en) * | 2019-07-03 | 2019-10-11 | 中山大学 | Location of mobile users prediction technique based on factor graph model |
Non-Patent Citations (1)
| Title |
|---|
| 基于神经网络的车辆轨迹预测算法的研究与实现;朱自升;《中国优秀硕士学位论文全文库》;20190215;正文第7-21页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110825833A (en) | 2020-02-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110825833B (en) | Method for predicting user moving track point | |
| CN107529651B (en) | Urban traffic passenger flow prediction method and equipment based on deep learning | |
| CN111836358B (en) | Positioning method, electronic device, and computer-readable storage medium | |
| CN110414732B (en) | Travel future trajectory prediction method and device, storage medium and electronic equipment | |
| CN111464938B (en) | Positioning method, positioning device, electronic equipment and computer readable storage medium | |
| CN111985710A (en) | Bus passenger trip station prediction method, storage medium and server | |
| CN108307301B (en) | Indoor positioning method based on RSSI ranging and track similarity | |
| CN105120433A (en) | WLAN indoor positioning method based on continuous sampling and fuzzy clustering | |
| CN111417066B (en) | Positioning method, positioning device, electronic equipment and computer readable storage medium | |
| CN111144452B (en) | Mobile user trip chain extraction method based on signaling data and clustering algorithm | |
| CN106568445A (en) | Indoor track prediction method based on bidirectional circulation neural network | |
| CN116186358B (en) | Depth track clustering method, system and storage medium | |
| CN111024098A (en) | Motor vehicle path fitting algorithm based on low-sampling data | |
| Niu et al. | RecNet: A convolutional network for efficient radiomap reconstruction | |
| CN115374375A (en) | Track clustering method, system and storage medium based on deep learning | |
| Uccellari et al. | On the use of support vector machines for the prediction of propagation losses in smart metering systems | |
| US20200271769A1 (en) | Method for evaluating positioning parameters and system | |
| CN112469116A (en) | Positioning method, positioning device, electronic equipment and computer readable storage medium | |
| CN103412963A (en) | Optical remote sensing satellite data obtaining method | |
| CN114758364B (en) | Industrial Internet of things scene fusion positioning method and system based on deep learning | |
| CN114595770B (en) | Long time sequence prediction method for ship track | |
| CN115457468A (en) | Intelligent livestock monitoring method and system for large grassland | |
| CN115495661A (en) | Self-adaptive interest point recommendation method based on long-term and short-term preference of user | |
| CN111695404B (en) | Pedestrian falling detection method and device, electronic equipment and storage medium | |
| ENETWILD‐consortium et al. | Update of model for wild boar abundance based on hunting yield and first models based on occurrence for wild ruminants at European scale |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |